The proliferation of antibiotic-resistant pathogens has spurred the use of drug combinations to maintain clinical efficacy and combat the evolution of resistance. However, the effect of multi- drug treatments on the rate of long term evolution of resistance is poorly understood. The phenotypic and genotypic evolutionary paths towards multi-drug resistance are complicated by the possibility of adaptation through different mutational steps, each of which may increase resistance either to an individual drug or to multiple drugs. Mutations can also affect the interactions between drugs, e.g. removing synergy among the drugs in a combination treatment. These types of adaptive steps often can represent tradeoffs between drug survival and growth rate. Tracking the adaptive evolutionary pathways a bacterial population follows towards multi-drug resistance has been difficult, since selection in fixed drug environments typically allows us to observe only a single adaptive step. A unique continuous culture device that we have developed, the 'morbidostat', automatically tunes drug concentration to maintain a fixed level of inhibition of evolving microbial populations; as bacteri become resistant, drug concentration automatically increases, thus allowing us to track evolution of resistance through multiple adaptive steps. Coupled with whole-genome sequencing and high-throughput growth and resistance phenotyping, this device will enable us to characterize the genotypic and phenotypic pathways of replicate Escherichia coli populations evolving high level of resistance to individual drugs (Aim 1), and to test how adaptation is affected and possibly slowed down by mixed drug combinations and alternating drug treatments (Aim 2). To systematically explore tradeoffs between growth and resistance, we will use a robust drug gradient assay to screen genomic libraries of gene deletion and over-expression strains for increased drug resistance and evaluate whether this increased resistance comes at the cost of decreased growth rate (Aim 3). Finally, we will identify the actual evolutionary trajectories towards multi-drug resistance taken by a pathogen in clinical settings, focusing on a small outbreak of Burkholderia dolosa in cystic fibrosis patients, where a single infecting strain has evolved in parallel in multiple individuals (Aim 4). Whole-genome sequencing allows reconstruction of the phylogeny of strains isolated from patients over time, and permits the identification of individual mutation events and their order of appearance. We will correlate these genetic changes with phenotypic measures of antibiotic resistance, and identify mutations that lead to drug resistance and their order of appearance. In summary, our proposed research provides a quantitative, systematic approach to understand the genotypic and phenotypic constraints that govern evolutionary pathways towards multi-drug resistance, both in the lab and during long-term infection of people. Our goal for the application of this research is to help design drug regimes that better prevent the emergence of resistance.